Caregiving and Work: The Relationship between Labor Market Attachment and Parental Caregiving

Published: 2017


There has been much concern over the provision of long-term care and the stresses it imposes on the family members who provide that care. However, despite the importance of this issue, it has been difficult to assess a causal relationship between caregiving and work. A chief concern is that those with weaker attachments to the labor force may be more willing to provide care—inducing a negative correlation when caregiving itself does not negatively affect employment. In this study we draw on 20 years of data from the Health and Retirement Study to examine anew the relationship between parental caregiving and work. We use two alternative identification strategies: First, we exploit the multiple observations per person existing in our data to estimate a fixed effects model for the relationship between caregiving and work. Second, we use unique data from the Social Security Administration on earnings histories to control for a woman’s labor market behavior long before the potential need to provide care. We find evidence that caregivers have at least a strong, and by some measures a stronger, relationship to the labor market than non-caregivers. Rather than labor force attachment, the provision of care appears to be driven primarily by parental need and by the availability of alternative caregivers, particularly sisters. However, we also find that caregiving has negative long-term effects on employment and earnings and can thus be detrimental to the financial well-being of caregivers.

Key Findings

    • Contrary to expectations, the women who are observed to provide care to a parent at some point during this survey window are not those with weaker attachments to the labor force.
    • Female caregivers have more schooling, are more likely to work full time, and have greater earnings than noncaregivers. They also have higher expected Social Security benefits. By contrast, we do not find similar patterns for those who provide care to parents-in-law.
    • Caregiving reduces the probability of work and hours worked in a simple regression, in a fixed-effects regression, and when controlling for work history using Social Security Administration records. However, we do not find a significant effect of caregiving on annual earnings.
    • At the end of the 20-year period, those women who provided care at some point had significantly lower annual earnings in 2010 than those who did not provide care, and this effect held both unconditionally and conditional on being employed.


Fahle, Sean and Kathleen McGarry. 2017. “Caregiving and Work: The Relationship between Labor Market Attachment and Parental Caregiving.” Ann Arbor, MI. University of Michigan Retirement Research Center (MRRC) Working Paper, WP 2017-356.